The problem is, it often takes *many* years before the errors and outright lies are found. In the meantime, the information is printed in millions of textbooks.
Why does it take so long for scientists to find the errors?
Depends on the field. Some disciplines have only one or two experts within a very narrow subfield, and they, as in the German paleontologist, can apparently bluff for a while. Other fields are more active, and have a lot more researchers within a specific subfield. Things can't be missed as easily there. Look at cold fusion, as an example. A number of other labs tried to duplicate the experiment within weeks, and within a few months it was becoming clear there was a problem.
Paleontology is a little different; it is not experimental and there are often very few people within an area/subfield. But there is no way for a mistake or fraud to last too long. The most famous one, Piltdown Man, was distrusted by many from the beginning. Only the British really liked it. By the time the South African fossils came on line it was clear one or the other was wrong, and so many folks just ignored Piltdown. A few, including Friedrichs and Weidenreich both, by about 1932, published research suggesting the lower jaws and molars were that of an orang (E.A. Hooton, Up from the Ape, revised edition; The MacMillan Co., 1946). This is what a 1946 textbook shows, several years before the claims for Piltdown were completely falsified.
With the internet, communication is much faster, and it takes a very narrow subspecialty, such as this German was in, for fraud to last very long at all.
So I hope this answers your question.